Hindawi Publishing CorporationCardiology Research and PracticeVolume 2013, Article ID 196506, 7 pageshttp://dx.doi.org/10.1155/2013/196506

Clinical StudyHeart Rate Variability in Children withTricyclic Antidepressant Intoxication

Ener Cagri Dinleyici,1,2 Zubeyir Kilic,3 Sabiha Sahin,1 Rabia Tutuncu-Toker,4

Makbule Eren,4 Zeynel Abidin Yargic,4 Pelin Kosger,3 and Birsen Ucar3

1 Department of Pediatric Intensive Care, Faculty of Medicine, Eskisehir Osmangazi University, 26480 Eskisehir, Turkey2 Pediatric Intensive Care Unit, Faculty of Medicine, Eskisehir Osmangazi University, 26480 Eskisehir, Turkey3 Department of Pediatric Cardiology, Faculty of Medicine, Eskisehir Osmangazi University, 26480 Eskisehir, Turkey4Department of Pediatrics, Faculty of Medicine, Eskisehir Osmangazi University, 26480 Eskisehir, Turkey

Correspondence should be addressed to Ener Cagri Dinleyici; timboothtr@yahoo.com

Received 15 December 2012; Revised 5 February 2013; Accepted 13 February 2013

Academic Editor: Gavin W. Lambert

Copyright © 2013 Ener Cagri Dinleyici et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

The aim of this study was to evaluate HRV in children requiring intensive care unit stays due to TCA poisoning between March2009 and July 2010. In the time-domain nonspectral evaluation, the SDNN (𝑃 < 0.001), SDNNi (𝑃 < 0.05), RMSDD (𝑃 < 0.01), andpNN50 (𝑃 < 0.01) were found to be significantly lower in the TCA intoxication group. The spectral analysis of the data recordedduring the first 5 minutes after intensive care unit admission showed that the values of the nLF (𝑃 < 0.05) and the LF/HF ratio(𝑃 = 0.001) were significantly higher in the TCA intoxication group, while the nHF (𝑃 = 0.001) values were significantly lower.The frequency-domain spectral analysis of the data recorded during the last 5 minutes showed a lower nHF (𝑃 = 0.001) in theTCA intoxication group than in the controls, and the LF/HF ratio was significantly higher (𝑃 < 0.05) in the intoxication group.The LF/HF ratio was higher in the seven children with seizures (𝑃 < 0.001). These findings provided us with a starting point forthe value of HRV analysis in determining the risk of arrhythmia and convulsion in TCA poisoning patients. HRV can be used as anoninvasive testing method in determining the treatment and prognosis of TCA poisoning patients.

1. Introduction

Poisoning is a common and important cause ofmorbidity andmortality worldwide, especially during childhood.More than50% of patients reported to poison control centers are adoles-cents or less than five years of age [1]. Among the substancesthat cause poisoning, drugs have been reported as the mostcommon factor. In Turkey, tricyclic antidepressants (TCA)occupy an important place among the causes of poison-ing, and the incidence of such events has increased over thelast decade [2]. TCAs (amitriptyline, nortriptyline, clomi-pramine, desipramine, imipramine, doxepinm, and protri-ptyline) are frequently used for the treatment of depression,chronic pain syndrome, school phobia, hyperkinesias, noc-turnal enuresis, and attention deficit-hyperactivity disorder[3].

TCA-related poisoning influences the autonomic nervoussystem, central nervous system (CNS), and cardiovascularsystem (CVS), and clinical signs appear within 6–8 hours.There is no laboratory test to precisely diagnose TCA-relatedpoisoning, and it is impossible to predict the prognosis. Thesemiquantitative enzyme immunoassay technique may beuseful for measuring TCA levels; however, there is no corre-lation between drug levels and clinical findings [4]. Themostsevere life-threatening condition in TCA-related poisoningis the dysrhythmias that occur due to the quinidine-likeeffects on the myocardial tissue. Electrocardiogram (ECG)findings vary [5], and Boehnert and Lovejoy [6] have shownan association between a QRS duration >100ms and the riskof arrhythmia or seizures.

Heart rate variability (HRV) can be described as instan-taneous differences in sinus rate over time or heart rate

2 Cardiology Research and Practice

fluctuations in the mean heart rate [7, 8]. Heart rate variabil-ity, which can be analyzed using ECG recordings, is a nonin-vasive method to assess cardiac autonomic activity. Changescan occur in heart rate that are related to autonomic tone dueto exercise, physical and mental stress, and respiratory andmetabolic reasons. A close relationship between increasedsympathetic activity or decreased parasympathetic activityand a tendency to fatal arrhythmia has been demonstrated[7]. HRV analysis that reflects the sympathetic and parasym-pathetic balance is used as a measure of cardiac autonomictone. Generally, 24-hour-long sequences should be obtainedto get reliable results for time-domain measures, and short-term recordings are needed to analyze the frequency-domainparameters. Studies regarding HRV in children are limitedcompared with the number of adult studies. HRV analysisseems to be promising in the study of the cardiovascularresponses to autonomic tone changes in obese children [9,10]. HRV analyses have been studied in children with beta-thalassemia [11, 12], epilepsy [13], recurrent neurocardiogenicsyncope [14], and obstructive sleep apnea [15, 16].

HRV plays an important role in describing fatal ornear-fatal arrhythmias [17] because increases in sympatheticactivity can cause severe arrhythmias and sudden death [18].Recently, a meta-analysis showed that the therapeutic useof TCAs caused a considerable decrease in HRV [19]. Therehave been analyses showingHRVchanges due toTCA-relatedpoisoning in adults, and differences in HRV have been sug-gested to be predictors for ventricular arrhythmias in an adultpatient [20, 21]. TCAs can be ingested accidentally or volun-tarily by children or therapeutic mistakes can occur. To ourknowledge, there have been no studies on the effects of TCA-related poisoning on HRV. In the present study, we aimed todetermine the effect of TCA poisoning on heart rate variabil-ity and to assess the effects of these results on clinical findings.

2. Materials and Methods

This prospective study was planned to involve children withisolated cases of TCA intoxication who had been referred tothe Pediatric EmergencyDepartment of EskisehirOsmangaziUniversity. We planned to enroll 20 children because ourprevious records showed that approximately 20 children hadbeen admitted to our clinic for TCA poisoning during thelast two years. For allocated randomization, 20 age- andsex-matched healthy children were included as a controlgroup. The children in the control group had no underlyingdiseases andwere admitted to our clinic for routine checkups.Holter analyses of the control children were recorded in theirhomes to avoid hospital stress. Essential equipment for theHRV analysis was provided within the scope of the EskisehirOsmangazi University Scientific Research Project (projectnumber: 200811022).The study protocol was approved by thelocal Ethical Committee of Eskisehir Osmangazi University,and the parents gave informed consent for the study.

TCA intoxication was diagnosed at >300 ng/mL of TCAas measured using semiquantitative enzyme immunoassaymethod in children who were referred to our clinic with ahistory or clinical findings of poisoning. The exclusion crite-ria were a history of drug intake apart from TCA, a different

drug toxicity according to the laboratory measurements, thepresence of chronic underlying diseases before poisoning,and the presence of obesity according to anthropometricmeasurements.The clinical findings for the patientswhowereincluded in the study were recorded during their PICU stay.Standard 12-lead electrocardiography and ambulatory bloodpressure measurements were also recorded. To perform theHRV analysis, HRV recordings were scheduled during thefirst half an hour after PICU admission in all of the cases.Recordings for a period of 24 hours that included the valuesfor an entire day and night were completed for both the studyand control groups.The parameters of the HRV analysis werecalculated in compliance with the recommendations of theAmerican Heart Association and the European Society ofCardiology, and the following parameters were evaluated inthe HRV analysis [7].

2.1. Time-Domain (Nonspectral) Analysis. This analysis wasbased on the assessment of the intervals between normalbeats on 24-hour ECG recordings. During the statisticalanalysis, generally all of the QRS complexes, the durationbetween consecutive QRS complexes (NN interval), or theinstantaneous heart rates during continuous ECG recordingsare determined [7]. We recorded the following time-domainindices for our HRV analysis: the standard deviation of all ofthe normal sinus RR intervals over 24 hours (ms) (SDNN),the percentage of successive normal sinus RR intervalsexceeding 50ms (%) (pNN50), and the root mean squareof the successive normal sinus RR interval differences (ms)(rMSSD). The SDANN (ms) and SDNN indices (ms) werealso recorded in all of the studied children.

2.2. Frequency-Domain (Spectral) Measures. Frequency-domain measures give information about how the power isdistributed as a function of frequency. We planned to useparametrical methods.The advantages of parametrical meth-ods are as follows: smoother spectral components that canbe distinguished as independent from preselected frequencybands, easy postprocessing of the spectrumwith an automaticcalculation of low- and high-frequency power componentsand an easy identification of the central frequency of eachcomponent, and accurate estimation even on a small numberof samples. We analyzed two different frequency domaincalculations during the first and last 5 minutes of the 24-hourrecordings in both the patient and control groups. The fre-quency domain indices were the low-frequency component(LF) (frequency range: 0.04–0.15Hz), the high-frequencycomponent (HF) (frequency range: 0.15–0.40Hz), and thevery low frequency component (VLF) (frequency range:0.003–0.04Hz) [7, 8]. Three values were noted for the firstand last 5-minute recordings. A normalized unit (nu) repre-sented the relative value of the LF or HF power componentsin proportion to the total power minus the VLF power. Thenormalizationminimized the effect of changes in total poweron the values of LF andHF.Theunit of the power componentsof theVLF, LF, andHFwasms2, but the unit of the normalizedLF (nLF) and normalized HF (nHF) was a normalized unit(nu). We used the nLF and nHF in the statistics. Inaddition, the LF/HF ratio was used to evaluate the balance

Cardiology Research and Practice 3

between the sympathetic and parasympathetic activities[7, 8].

Statistical analysis of study data was performed usingSPSS 16.0 Software (Chicago, IL). The Mann-Whitney 𝑈 testwas used for comparisons. Comparisons over time wereperformed using the Wilcoxon test, and correlations wereevaluated using the Spearman correlation test. A 𝑃 < 0.05was considered statistically significant.

3. Results

Twenty children (nine boys and eleven girls) between the agesof 3 and 16 years old following admittance to the PICU withTCA intoxication (TCA level > 300 ng/mL) and 20 healthychildren (10 boys and 10 girls) were included in the studybetween March 2009 and July 2010. Sixteen of the patientspresenting with TCA intoxication had ingested the drug forthe purpose of suicide, and 4 had ingested the drug accidently.The median time to arrive in the emergency department was2.4 hours (range 0.5–4 hours). The total serum TCA levelranged from 380 to 2000 ng/mL (mean 1116 ± 635 ng/mL).The most commonly observed clinical sign was the loss ofconsciousness (𝑛 = 16, 80%). Seizure was observed in 7cases (35%).Themost frequent cardiovascular sign was sinustachycardia (85%, 𝑛 = 17). The mean heart rate was 148 ±36 beats/minute (range 66–224 beats/min).Themean systolicblood pressure was 115mmHg (range 62–160mmHg), andthe mean diastolic blood pressure was 58mmHg (range 36–92mmHg) during the blood pressure monitoring. The meansystolic and diastolic blood pressures were normal in thecontrol group. Although there were some differences in bloodpressure in the TCA intoxication group, no statistical differ-ence was found between the groups (𝑃 > 0.05). AbnormalECG results were found in 12 of the cases (60%). QTc pro-longation was observed in 5 cases, PR interval prolongationwas observed in 3 cases, and QRS interval prolongation wasobserved in 4 cases (Table 1). ST-T alteration was observed inonly one case. No Brugada sign was observed. There was apositive correlation between serum TCA levels and the QRSinterval (𝑟 = 0.684; 𝑃 < 0.01). The mean age was higher inpatients with positive ECG findings compared with patientswith normal ECG results (median 14.6 years versus 11.2 years,resp., 𝑃 < 0.05). The TCA levels in patients with positiveECG findings were significantly higher than in patientswithout positive ECG findings (1523 ± 571 ng/mL versus 619± 211 ng/mL; 𝑃 = 0.001). The duration of the PICU staywas markedly longer in patients with positive ECG findingscompared with patients without ECG findings (median 5.4days versus 2.4 days; 𝑃 < 0.01). All of the cases weretreated using activated charcoal and alkalization. No case wastreated using physostigmine, plasmapheresis, or hemoperfu-sion. One patient was mechanically ventilated for a period of24 hours. The median duration of the intensive care unit staywas 4.7 days (range 1–9 days), and total duration of hospitalstay was 5.4 days (range 3–12 days). All of the patients weresuccessfully treated with these treatments and discharged.

3.1. HRV Analysis. The time-domain nonspectral parame-ters (SDNN, SDANN, RMSDD, SDNN index, NN50, and

Table 1: Clinical and laboratory findings of children with TCAintoxication.

Children with TCA intoxication(𝑛 = 20)Age 3–16 yearsGender 9 boys, 11 girls

TCA ingestionPurpose of suicide

(𝑛 = 16)Accidently (𝑛 = 4)

Serum TCA levels 1116 ± 635 g/mL(380–2000 ng/mL)

Loss of consciousness 16/20Seizures 7/20Median heart rate 148 (66–224)Median systolic blood pressure(mmHg) 115 (62–160)

Median diastolic blood pressure(mmHg) 58 (36–92)

Abnormal ECG results 12 (20)

Prolonged PR interval 3 cases (0.24, 0.28, 0.24second)

Prolonged QTc interval 5 cases (0.48, 0.52, 0.48,0.47, 0.51)

Prolonged QRS interval 4 cases (>100ms for all 4cases)

PNN50) recorded over the course of 24 hours were evaluatedin all of the patients and controls, and the results are shownas the median and 95% CI in Table 2. The median SDNN,RMSDD, and SDNN index values were lower in the TCAgroup than in the controls (𝑃 < 0.001,𝑃 < 0.05, and𝑃 < 0.01,resp.). In the TCA intoxication group, the pNN50 values werelower (𝑃 < 0.01 for both). The SDNN (𝑃 < 0.001), RMSDD(𝑃 < 0.01), SDDNi (𝑃 < 0.01), and pNN50 (𝑃 < 0.01) valueswere lower in the TCA intoxication group among patientswith ECG findings compared with patients without positiveECG findings.

There was no significant difference in the VLF valuesduring the first 5 minutes of the intensive care unit staybetween the TCA intoxication and control groups (𝑃 > 0.05).The nLF was markedly higher in the TCA intoxication groupcompared with the control group (𝑃 < 0.05), and the nHFwasmarkedly lower in the TCA intoxication group comparedwith the control group (𝑃 = 0.001). Also, the LF/HF ratiowas calculated to bemarkedly higher in the TCA intoxicationgroup compared with the control group (𝑃 = 0.001, Table 2and Figure 1).

The frequency-domain spectral analysis was repeated forthe last 5 minutes during the 24-hour HRV recordings inboth groups. Similarly to the first 5 minutes, the VLF valueswere found to be similar for both groups. While there wasno difference in the nLF values, the nHF was lower in theTCA intoxication group compared with the patients duringthe first 5 minutes (𝑃 = 0.001). The LF/HF ratio at the end ofthe 24 hours in the TCA intoxication group was higher thanthat in control group (𝑃 < 0.05).

4 Cardiology Research and Practice

Table 2: Twenty-four-hour time-domain (non-spectral) and 5-minute spectral analysis results by study group.

TCAintoxication(𝑛 = 20)

Control(𝑛 = 20) 𝑃

Mean heartrate (perminute)

148(66–224)

78(55–102) <0.05

SDNN (ms) 104(94–117)

146(134–170) 0.000

SDANN (ms) 27(25–44)

31(26–39) 𝑃 > 0.05

RMSDD (ms) 46(41–64)

65(57–102) 0.049

SDNN index 60(51–70)

81(72–96) 0.002

pNN50 (%) 11.3(8.9–20.9)

24.6(20.8–33.4) 0.005

Frequency-domain (spectral) analysis

VLF (ms2) 772(271–3873)

806(698–1046) 𝑃 > 0.05

nLF (nu) 27(24–35)

17(11–24) 0.012

nHF (nu) 6.9(5.4–8.7)

12.1(10.7–22) 0.001

LF/HF 4.6(1.1–14.4)

0.96(0.8–2) 0.001

∗Values are given as median (95% CI).

TCA Control0

1

2

3

4

5

6

7

8

9

10

LF/H

F

Figure 1: Comparison of the LF/HF ratio within the first 5 minutesbetween children with TCA intoxication and controls.

On comparing the first and last 5-minute recordings ofthe 20 children with TCA intoxication, no differences weredetermined for the VLF, nHF, or LF/HF ratio; however,the nLF value for the patients whose nLF value was highercontinued to remain high (𝑃 < 0.05).

0

1

2

3

4

5

6

7

8

9

10

LF/H

F

Seizure (+) Seizure (−)

Figure 2: Comparison of the LF/HF ratio during the first 5 minutesof the HRV analysis according to the presence of seizures in patientswith TCA intoxication.

There were no significant differences in the time-domainnonspectral analysis parameters between cases with or with-out seizures; however, the nLF value during the first 5minuteswas higher (though not statistically significantly higher), andthe nHF value was lower (𝑃 < 0.05) in the group whohad seizures. Also, the LF/HF ratio was found to be signifi-cantly higher in patients who had seizures compared withpatientswithout seizures (median 6.95 versus 3.79,𝑃 < 0.001)(Figure 2).

When we checked the association between the serumTCA levels and theHRVparameters in the patients with TCAintoxication, we found that the TCA levels showed a negativecorrelation with the SDNN (𝑟 = −0.883, 𝑃 < 0.001), RMSDD(𝑟 = −0.564, 𝑃 < 0.01), SDNNi (𝑟 = −0.469, 𝑃 < 0.05),SDSD (𝑟 = −0.564, 𝑃 < 0.01), and pNN50 (𝑟 = −0.601,𝑃 < 0.01). We found no significant association between theTCA levels and the spectral analysis parameters (nLF, NHF,LF/HF) during the first and last 5 minutes of recording.

4. Discussion

Tricyclic antidepressant intoxication can lead cardiovascular,respiratory, and central nervous system signs. The cardiovas-cular effects of TCA intoxication include PR or QT inter-val changes, QRS complex enlargement, and ventricular orsupraventricular tachyarrhythmia, which originates basicallyfrom the quinidine-like effects of these drugs on heart tissue[22]. The most frequently observed dysrhythmia is sinustachycardia, which was observed in our study, and wide com-plex tachycardia (supraventricular or ventricular) remainsone of the characteristic complications of TCA intoxication[23]. Leonard et al. [24] similarly have described tachycardiaas the most frequent cardiac sign, and they have shown thatsinus tachycardia accompanies PR, QRS, and QTc changes inchildren similarly to adults; however, ECG findings have no

Cardiology Research and Practice 5

diagnostic value for evaluating TCA intoxication in childrenand adults [23]. It has been concluded that increases >0.10 secin the QRS interval length, >3mm in the height of aVR Ramplitude, and the abnormalities on the last 40ms of a QRSaxis on the frontal axis are significant for convulsion andventricular arrhythmia in cases with TCA intoxication [6].But there is no predictive value for the assessment of theterminal 40ms in children as in adults [25].𝐴 > 0.10 sec QRSduration in extremity derivations has been associated withconvulsion, and 𝐴 > 0.16 sec QRS interval has been associ-ated with ventricular arrhythmia. Citak et al. [26] reportedcardiac signs in almost half of all patients with convulsionamong their cases. In our study, 7 cases (35%) had seizures;however, no associationwas foundbetweenECGchanges andseizures. There were also no associations between the TCAlevels and the presence of conduction disorders or seriouscardiac dysrhythmias [27]. In our study, a positive correlationwas found between TCA levels and QRS duration, and theserum TCA levels in children with ECG findings were higherin our study. Our results indicated that ECG findings canbe used to evaluate TCA intoxication; however, our patientnumber was not sufficient to show evidence of an associationbetween the ECG findings and serum TCA levels.

HRV analysis has been used to diagnose, follow up, ordetermine the prognosis of many disorders and is presumedto be safe as a method to quantitatively assess changes in theautonomic nervous system and determine the cardiovascularresponse to these changes. In many conditions, the balancebetween the sympathetic and parasympathetic systems isaffected, and cardiac autonomic function disorders appear[28, 29]. Abnormalities on HRV as a marker of cardiac auto-nomic control are presumed to be risk factors for arrhythmia,and it has been shown that they can be used as reliablenoninvasive markers of the presence of an underlying disease[30–33]. In our study, HRV abnormalities were clearly shownin children with TCA intoxication during the first 24 hoursof their admission. With regard to time-domain analysis,SDNN reflected the overall HRV and the SDSS and RMSSDreflected the vagal tone. In addition, the RMSSD and pNN50parameters indicated high-frequency variations in heart rateindependently from diurnal or other changes [7]. In ourstudy, the significantly decreased SDNN, RMSDD, and PN50values in children with TCA intoxication indicated the effectof TCAs on the cardiac autonomic system. Also, the negativecorrelation between TCA levels and SDNN values suggestedthat time-domain measures can be used as a marker to deter-mine drug levels in cases of TCA intoxication. Decreasedtime-domain parameters, particularly SDNN, gave an ideaabout the association between the HRV results and ECGfindings. Bjelakovic et al. [33] found a marked decrease inthe SDNN and SDANN levels in adults with exercise-inducedventricular arrhythmia, and they suggested that this decreasemight be related to the direct effect of the sympathetic systemonmyocardial electrical activity. Changes in the RMSDD andpNN50 components also indicated that the autonomic systemwas affected by TCA poisoning via the vagus nerve.

The spectral analysis, including the short-term HRVrecordings ranging between 2 to 5 minutes, is accepted as anoninvasivemarker for the sympathetic and parasympathetic

systems. In this way, heart rate signals are separated byfrequency and intensity. Among these frequency bands, theLH and HF and their ratio (LF/HF) are frequently used[7]. The HF is thought to be the main marker of parasym-pathetic activity, but the LF reflects both the sympatheticand parasympathetic activity. Numerous investigators haveemphasized that an increased LF/HF ratio shows sympatheticactivity dominance. We evaluated the VLF, nLF, nHF, andLF/HF ratio during the first and last 5 minutes of the 24-hourHRV recordings. The nLF, nHF, and LF/HF ratio recordedat the intensive care unit were higher during the first 5minutes of HRV recording in patients with TCA intoxication.During the last 5 minutes at the end of 24-hour record,the nLF levels were similar, the nHF level was lower, andthe LF/HF remained high. One of the main determinantsof parasympathetic activity is the high-frequency compo-nent (HF); thus, parasympathetic activity continued to bedepressed in our patients after 24 hours of intoxication.Theseresults indicated that the first 24-hour period is critical inTCA intoxication and that the changes that occurred at theend of the 24 hours may be explained by treatment for andelimination of the drug. While there were no differences inthe nonspectral analysis parameters between patients withand without convulsions, the nLF and the LF/HF ratio werehigher and the nHF was lower in patients with seizures.Frequency analysis in the first 5 minutes (especially LF/HFratio) could be predictive of seizures and be an indication foranticonvulsant treatment. More comprehensive new studiesare needed to evaluate the predictive value of the LF/HF ratioat admission for different treatment modalities.

Eisenhofer et al. [34] performed an experimental studyfor evaluation of the effects of desipramine (member ofTCAs) on sympathetic nerves and they showed that desipra-mine has exerted differential central and peripheral effects(reduced sympathetic nerve outflow and diminished nore-pinephrine uptake). In humans, while desipramine dimin-ished to norepinephrine uptake, it increased cardiac norepin-ephrine spillover by 25% [35]. There have been publica-tions relating to HRV changes in adults and children whoreceived TCA therapeutically [19, 36–39]. Srinivasan et al.[39] reported a marked decrease in the HF and an increasedLF/HF ratio in the standing position, indicating the doseeffects of therapeutic TCA on the cardiac autonomic sys-tem in children. Regarding TCA intoxication, antidepres-sant drugs are among themost commonmeans of attemptingsuicide, and they are a serious risk factor for arrhythmias.Numerous studies have been performed to determine thedevelopment, type, and the severity of arrhythmia, and nostandard indicator has yet been found [20]. It has been shownthat HRV is markedly suppressed after imipramine andamitriptyline intake, and even if the EGC is normal, there isan association between the development of ventricular arrhy-thmia and increased LF/HF ratios [21]. Waring et al. [20]compared the results of HRV analysis in patients with antide-pressant drug intoxication, and they found markedly pro-longed QT and QTc intervals in patients who had receivedantidepressants, with a slight decrease in the HF and aslight increase in the LF. Djonlagic et al. [21] suggestedthat HRV is useful to assess the autonomic status of the

6 Cardiology Research and Practice

cardiac system during the early and late terms of TCA intoxi-cation. In particular, the cardiac autonomic system status canbe obtained using HRV, and the risk of severe arrhythmiacan be described in high-risk patients after TCA intoxication.Djonalgic et al. [21] reported the results of HRV in a 28-year-old patient who was referred with signs of TCA intoxica-tion. In this case, the QRS complexes were wider than 150ms,and the QTc interval was prolonged. The total power andLF and HF values were markedly suppressed, and the HRVdepression continued during the second day without path-ology on ECG. A markedly increased LF/HF ratio wasnoted before the development of ventricular arrhythmia.Theheart rate and rhythm were markedly under the control ofthe autonomic nervous system. Thus, increased sympatheticactivity in our study could result in clinical signs, particularlyincreased heart rate. It can be observed that parasympatheticsuppression disappears, sympathetic activity becomes dom-inant, heart rate increases, and serious tachycardia occursduring TCA intoxication.

Power spectral analysis of HRV has been used frequentlyto assess cardiac autonomic function in different clinical con-dition; however, Goldstein et al. [40] highlighted that therelationship of LF power of HRV to cardiac sympathetic tonehas been unclear. Manipulations and drugs that change LF/HF may do so not by affecting cardiac autonomic outflowsdirectly but by affecting modulation of those outflows by bar-oreflexes. Watson et al. [41] showed that the increased card-iac sympathetic nerve activity is not dependent on depressedarterial baroreflex control of cardiac sympathetic nerve activ-ity in experimental study about heart failure. For example,heart failure is characterized by activation of the sympath-etic nervous system as demonstrated in patients by increasesin circulating norepinephrine, total and regional norepine-phrine spillover, and muscle sympathetic nerve activity. Forthis reason other determinant affecting HRV analysis includ-ing underlying disease, interventions, and drugs should beextensively evaluated during this analysis. Kingwell et al. [42]showed that the combination of the cardiac norepinephrinespillover technique and HRV will allow a more comprehen-sive assessment of both neuronal and postsynaptic aspects ofthe cardiac neuroeffector response. Our study, like previousclinical HRV studies, demonstrated some HRV abnormal-ities; however, extensive evaluation of cardiac autonomicactivity needs includes other potential contributors.

In conclusion, autonomic changes that affect the heartrate may particularly cause ECG changes and arrhythmiasthat may develop in cases with TCA intoxication.Thus, treat-ments and antidotes the mechanisms of cause an increase invagal activity may be effective in TCA intoxication. Short-term heart rate variability recordings may be a guide, espe-cially for rhythm changes and as a predictor for seizures. Weconcluded that HRV could be used in the future as a nonin-vasive evaluationmethod for early diagnosis, risk determina-tion, and determination of the effectiveness of treatment.

Acknowledgment

This study was supported by Eskisehir Osmangazi UniversityScientific Research Project (project number: 200811022).

References

[1] T. L. Litovitz, W. Klein-Schwartz, E. M. Caravati, J. Youniss, B.Crouch, and S. Lee, “1998 annual report of the American Asso-ciation of Poison Control Centers Toxic Exposure SurveillanceSystem,” American Journal of Emergency Medicine, vol. 17, no. 5,pp. 435–487, 1999.

[2] N. Andıran and F. Sarıkayalar, “Pattern of acute poisonings inchildren in Ankara: what has changed in twenty years?” TheTurkish Journal of Pediatrics, vol. 46, pp. 147–152, 2004.

[3] P. K.Gillman, “Tricyclic antidepressant pharmacology and ther-apeutic drug interactions updated,” British Journal of Pharma-cology, vol. 151, no. 6, pp. 737–748, 2007.

[4] G.W. Kerr, A. C.McGuffie, and S.Wilkie, “Tricyclic antidepres-sant overdose: a review,”EmergencyMedicine Journal, vol. 18, no.4, pp. 236–241, 2001.

[5] E.M. Caravati, “The electrocardiogram as a diagnostic discrim-inator for acute tricyclic antidepressant poisoning,” Journal ofToxicologyy, vol. 37, no. 1, pp. 113–115, 1999.

[6] M. T. Boehnert and F. H. Lovejoy, “Value of the QRS durationversus the serum drug level in predicting seizures and ventri-cular arrhythmias after an acute overdose of tricyclic antide-pressants,” The New England Journal of Medicine, vol. 313, no.8, pp. 474–479, 1985.

[7] Task Force of the European Society of Cardiology and theNorth American Society of Pacing and Electrophysiology,“Heart rate variability: standards ofmeasurement, physiologicalinterpretation, and clinical use,” European Heart Journal, vol. 17,pp. 354–381, 1996.

[8] L. C. M. Vanderlei, C. M. Pastre, R. A. Hoshi, T. D. de Carvalho,and M. F. de Godoy, “Basic notions of heart rate variability andits clinical applicability,”Brazilian Journal of Cardiovascular Sur-gery, vol. 24, no. 2, pp. 205–217, 2009.

[9] C. L. Kaufman, D. R. Kaiser, J. Steinberger, A. S. Kelly, and D.R. Dengel, “Relationships of cardiac autonomie function withmetabolic abnormalities in childhood obesity,” Obesity, vol. 15,no. 5, pp. 1164–1171, 2007.

[10] S. M. Rodrıguez-Colon, E. O. Bixler, X. Li, A. N. Vgontzas,and D. Liao, “Obesity is associated with impaired cardiac auto-nomic modulation in children,” International Journal of Pedi-atric Obesity, vol. 6, no. 2, pp. 128–134, 2011.

[11] W. Rutjanaprom, N. Kanlop, P. Charoenkwan et al., “Heart ratevariability in beta-thalassemia patients,” European Journal ofHaematology, vol. 83, no. 5, pp. 483–489, 2009.

[12] F. Kardelen, G. Tezcan, G. Akcurin, H. Ertug, and A. Yesilipek,“Heart rate variability in patients with thalassemia major,”Pediatric Cardiology, vol. 29, no. 5, pp. 935–939, 2008.

[13] O. Hallioglu, C. Okuyaz, E. Mert, and K. Makharoblidze,“Effects of antiepileptic drug therapy on heart rate variabilityin children with epilepsy,” Epilepsy Research, vol. 79, no. 1, pp.49–54, 2008.

[14] M. Akcaboy, S. Atalay, T. Ucar, and E. Tutar, “Heart rate vari-ability during asymptomatic periods in children with recurrentneurocardiogenic syncope,” The Turkish Journal of Pediatrics,vol. 53, no. 1, pp. 59–66, 2011.

[15] K. L. Kwok, T. C. Yung, D. K. Ng, C. H. Chan, W. F. Lau, andY. M. Fu, “Heart rate variability in childhood obstructive sleepapnea,” Pediatric Pulmonology, vol. 46, no. 3, pp. 205–210, 2011.

[16] H. V. Muzumdar, S. Sin, M. Nikova, G. Gates, D. Kim, and R.Arens, “Changes in heart rate variability after adenotonsillec-tomy in children with obstructive sleep apnea,” Chest, vol. 139,no. 5, pp. 1050–1059, 2011.

Cardiology Research and Practice 7

[17] H. V. Huikuri, M. J. P. Raatikainen, R. Moerch-Joergensen et al.,“Prediction of fatal or near-fatal cardiac arrhythmia events inpatients with depressed left ventricular function after an acutemyocardial infarction,” European Heart Journal, vol. 30, no. 6,pp. 689–698, 2009.

[18] M. T. La Rovere, G.D. Pinna, R.Maestri et al., “Short-termheartrate variability strongly predicts sudden cadiac death in chronicheart failure patients,” Circulation, vol. 107, no. 4, pp. 565–570,2003.

[19] A. H. Kemp, D. S. Quintana, M. A. Gray, K. L. Felmingham, K.Brown, and J.M.Gatt, “Impact of depression and antidepressanttreatment on heart rate variability: a review and meta-analysis,”Biological Psychiatry, vol. 67, no. 11, pp. 1067–1074, 2010.

[20] W. S. Waring, J. Y. Rhee, D. N. Bateman, G. E. Leggett, andH. Jamie, “Impaired heart rate variability and altered card-iac sympathovagal balance after antidepressant overdose,”Euro-pean Journal of Clinical Pharmacology, vol. 64, no. 11, pp. 1037–1041, 2008.

[21] I. Djonlagic, H. Djonlagic, T. Kibbel, S. Suefke, and C. Dodt,“Heart rate variability reveals risk of arrhythmias after intoxica-tion with antidepressants,” Intensive Care Medicine, vol. 33, no.1, pp. 200–202, 2007.

[22] P.Manikoth, R. Subramanyan, S.Menon, and S.M.AlKhusaiby,“A child with cardiac arrhythmia and convulsions,”The Lancet,vol. 354, no. 9195, p. 2046, 1999.

[23] T. G. Rosenbaum, M. Kou, and J. N. Love, “Are one or twodangerous? Tricyclic antidepressant exposure in toddlers,” Jour-nal of Emergency Medicine, vol. 28, no. 2, pp. 169–174, 2005.

[24] H. L. Leonard, M. C. Meyer, S. E. Swedo et al., “Electrocardio-graphic changes during desipramine and clomipramine treat-ment in children and adolescents,” Journal of the AmericanAcademy of Child and Adolescent Psychiatry, vol. 34, no. 11, pp.1460–1468, 1995.

[25] J. Sarko, “Antidepressants, old and new: a review of their adverseeffects and toxicity in overdose,” Emergency Medicine Clinics ofNorth America, vol. 18, no. 4, pp. 637–654, 2000.

[26] A. Citak, D. D. Soysal, R. Ucsel, M. Karabocuoglu, and N. Uzel,“Seizures associatedwith poisoning in children: tricyclic antide-pressant intoxication,” Pediatrics International, vol. 48, no. 6, pp.582–585, 2006.

[27] L. P. James and G. L. Kearns, “Cyclic antidepressant toxicity inchildren and adolescents,” Journal of Clinical Pharmacology, vol.35, no. 4, pp. 343–350, 1995.

[28] P. Palatini and S. Julius, “The role of cardiac autonomic functionin hypertension and cardiovascular disease,” Current Hyperten-sion Reports, vol. 11, no. 3, pp. 199–205, 2009.

[29] A. Haensel, P. J. Mills, R. A. Nelesen, M. G. Ziegler, and J. E.Dimsdale, “The relationship between heart rate variability andinflammatory markers in cardiovascular diseases,” Psychoneu-roendocrinology, vol. 33, no. 10, pp. 1305–1312, 2008.

[30] K. P. Anderson, “Sympathetic nervous system activity and ven-tricular tachyarrhythmias: recent advances,”Annals of Noninva-sive Electrocardiology, vol. 8, no. 1, pp. 75–89, 2003.

[31] A. Algra, J. G. P. Tijssen, J. R. T. C. Roelandt, J. Pool, and J. Lub-sen, “Heart rate variability from 24-hour electrocardiographyand the 2-year risk for sudden death,” Circulation, vol. 88, no. 1,pp. 180–185, 1993.

[32] A. F. Folino, G. Russo, B. Bauce, E. Mazzotti, and L. Daliento,“Autonomic profile and arrhythmic risk stratification aftersurgical repair of tetralogy of Fallot,” American Heart Journal,vol. 148, no. 6, pp. 985–989, 2004.

[33] B. Bjelakovic, S. Ilic, K. Chouliaras et al., “Heart rate vari-ability in children with exercise-induced idiopathic ventriculararrhythmias,” Pediatric Cardiology, vol. 31, no. 2, pp. 188–194,2010.

[34] G. Eisenhofer, T. Saigusa, M. D. Esler, H. S. Cox, J. A. Angus,and P. K. Dorward, “Central sympathoinhibition and peripheralneuronal uptake blockade after desipramine in rabbits,” Ameri-can Journal of Physiology, vol. 260, no. 4, part 2, pp. R824–R832,1991.

[35] M. D. Esler, G. Wallin, P. K. Dorward et al., “Effects of desi-pramine on sympathetic nerve firing and norepinephrine spill-over to plasma in humans,” American Journal of Physiology, vol.260, no. 4, part 2, pp. R817–R823, 1991.

[36] K. Udupa, T. N. Sathyaprabha, J. Thirthalli et al., “Alteration ofcardiac autonomic functions in patients with major depression:a study using heart rate variabilitymeasures,” Journal of AffectiveDisorders, vol. 100, no. 1–3, pp. 137–141, 2007.

[37] F. Lederbogen, C. Gernoth, B. Weber et al., “Antidepressivetreatmentwith amitriptyline and paroxetine: comparable effectson heart rate variability,” Journal of Clinical Psychopharmacol-ogy, vol. 21, no. 2, pp. 238–239, 2001.

[38] C. A. Arranto, C. Mueller, P. R. Hunziker, S. C. Maysch, andU. Eriksson, “Adverse cardiac events in ICU patients with pre-sumptive antidepressant overdose,” Swiss Medical Weekly, vol.133, no. 35-36, pp. 479–483, 2003.

[39] K. Srinivasan, M. V. Ashok, M. Vaz, and V. K. Yeragani, “Effectof imipramine on linear and nonlinear measures of heart ratevariability in children,” Pediatric Cardiology, vol. 25, no. 1, pp.20–25, 2004.

[40] D. S. Goldstein, O. Bentho, M. Y. Park, and Y. Sharabi, “Low-frequency power of heart rate variability is not a measure ofcardiac sympathetic tone but may be a measure of modulationof cardiac autonomic outflows by baroreflexes,” ExperimentalPhysiology, vol. 96, pp. 1255–1261, 2011.

[41] A. M. D. Watson, S. G. Hood, R. Ramchandra, R. M. McAllen,and C. N. May, “Increased cardiac sympathetic nerve activityin heart failure is not due to desensitization of the arterial baro-reflex,”American Journal of Physiology, vol. 293, no. 1, pp.H798–H804, 2007.

[42] B. A. Kingwell, J. M.Thompson, D. M. Kaye, G. A. McPherson,G. L. Jennings, and M. D. Esler, “Heart rate spectral analysis,cardiac norepinephrine spillover, andmuscle sympathetic nerveactivity during human sympathetic nervous activation andfailure,” Circulation, vol. 90, no. 1, pp. 234–240, 1994.